Precise truck parking method, device, and system in quayside container crane area

Patent metadata

Field	Value
Publication number	US-12043248-B2
Application number	US-202117360649-A
Country	US
Kind code	B2
Filing date	Jun 28, 2021
Priority date	Dec 26, 2018
Publication date	Jul 23, 2024
Grant date	Jul 23, 2024

Abstract

A method, a device, and a system for parking a truck accurately in a shore crane area are provided. A vehicle controller transmits a parking request for a truck to be parked. A main controller receives the parking request and acquires real-time point cloud data by scanning one or more lanes crossed by a shore crane using one or more LiDARs. The main controller clusters the real-time point cloud data to obtain a set of point clouds for the truck and applies an Iterative Closest Point algorithm to the set of point clouds and a vehicle point cloud model to obtain a real-time distance from the truck to a target parking space. The vehicle controller controls the truck to stop at the target parking space based on the real-time distance.

First claim

What is claimed is:

1. A method, comprising:
receiving a parking request for a vehicle to be parked from a vehicle controller;
acquiring real-time point cloud data by scanning one or more lanes crossed by a shore crane using one or more LiDARs in response to receiving the parking request;
clustering the real-time point cloud data to obtain a set of point clouds for the vehicle to be parked;
obtaining a real-time distance from the vehicle to be parked to a target parking space based on the set of point clouds for the vehicle to be parked and a vehicle point cloud model, comprising:
determining a vehicle model of the vehicle to be parked; and
selecting the vehicle point cloud model matching the vehicle model of the vehicle to be parked from a model library, wherein the model library comprises a plurality of vehicle point cloud models obtained by scanning vehicles of a plurality of different vehicle models that are parked at the target parking space using the LiDAR in advance; and

transmitting to the vehicle controller the real-time distance from the vehicle to be parked to the target parking space, such that the vehicle controller controls the vehicle to be parked to stop at the target parking space based on the real-time distance.

2. The method of claim 1 , wherein said obtaining the real-time distance from the vehicle to be parked to the target parking space based on the set of point clouds for the vehicle to be parked and the vehicle point cloud model further comprises:
applying an Iterative Closest Point (ICP) algorithm to the set of point clouds for the vehicle to be parked and the vehicle point cloud model matching the vehicle model of the vehicle to be parked.

3. The method of claim 1 , wherein said selecting the vehicle point cloud model matching the vehicle model of the vehicle to be parked from the model library comprises:
selecting, in response to determining that the vehicle to be parked carries a container, the vehicle point cloud model matching the vehicle model of the vehicle to be parked from a first model library, wherein the first model library comprises a plurality of vehicle point cloud models obtained by scanning vehicles of the plurality of different vehicle models that are parked at the target parking space and carrying containers using the LiDAR in advance; and
selecting, in response to determining that the vehicle to be parked carries no container, the vehicle point cloud model matching the vehicle model of the vehicle to be parked from a second model library, wherein the second model library comprises a plurality of vehicle point cloud models obtained by scanning vehicles of the plurality of different vehicle models that are parked at the target parking space and carrying no container using the LiDAR in advance.

4. The method of claim 1 , wherein said obtaining the real-time distance from the vehicle to be parked to the target parking space based on the set of point clouds for the vehicle to be parked and the vehicle point cloud model further comprises:
calculating a translation matrix from the set of point clouds for the vehicle to be parked to the vehicle point cloud model using an Iterative Closest Point (ICP) algorithm; and
obtaining the real-time distance from the vehicle to be parked to the target parking space based on the translation matrix.

5. The method of claim 4 , wherein said calculating the translation matrix from the set of point clouds for the vehicle to be parked to the vehicle point cloud model using the ICP algorithm comprises:
determining an initial translation matrix, the initial translation matrix being a matrix for translating an average center of a predetermined number of foremost points in a moving direction of the vehicle to be parked in the set of point clouds for the vehicle to be parked to an average center of the predetermined number of foremost points in the moving direction of the vehicle to be parked in the vehicle point cloud model, wherein coordinates of the average center are average values of coordinates of the predetermined number of points; and
performing iterative calculation on the set of point clouds for the vehicle to be parked and the vehicle point cloud model based on the initial translation matrix, to obtain the translation matrix from the set of point clouds for the vehicle to be parked to the vehicle point cloud model.

6. The method of claim 4 , wherein said calculating the translation matrix from the set of point clouds for the vehicle to be parked to the vehicle point cloud model using the ICP algorithm comprises:
parsing positioning data collected by a vehicle-mounted positioning device on the vehicle to be parked from the parking request;
determining an initial translation matrix, the initial translation matrix being a matrix for translating one or more points corresponding to the positioning data to one or more positioning points corresponding to the target parking space, the one or more positioning points being one or more points corresponding to positioning data obtained by the vehicle-mounted positioning device when the vehicle stops at the target parking space in a process for determining the vehicle point cloud model; and
performing iterative calculation on the set of point clouds for the vehicle to be parked and the vehicle point cloud model based on the initial translation matrix, to obtain the translation matrix from the set of point clouds for the vehicle to be parked to the vehicle point cloud model.

7. The method of claim 1 , wherein said receiving the parking request comprises:
receiving the parking request;
parsing a shore crane identification from the parking request;
matching the shore crane identification with a shore crane corresponding to itself; and
establishing a communication connection with the vehicle controller when the shore crane identification matches the shore crane.

8. The method of claim 1 , wherein said clustering the real-time point cloud data to obtain the set of point clouds for the vehicle to be parked comprises:
parsing positioning data collected by a vehicle-mounted positioning device on the vehicle to be parked from the parking request;
clustering the real-time point cloud data based on the positioning data to obtain the set of point clouds for the vehicle to be parked.

9. The method of claim 1 , wherein said clustering the real-time point cloud data to obtain the set of point clouds for the vehicle to be parked comprises:
parsing positioning data collected by a vehicle-mounted positioning device on the vehicle to be parked from the parking request;
extracting, from the real-time point cloud data, point cloud data corresponding to a position corresponding to the positioning data and an area within a predetermined length around the position; and
clustering the extracted point cloud data to obtain the set of point clouds for the vehicle to be parked.

10. The method of claim 1 , wherein said clustering the real-time point cloud data to obtain the set of point clouds for the vehicle to be parked comprises:
parsing a number of a lane where the vehicle to be parked is located from the parking request;
extracting, from the real-time point cloud data, point cloud data of the lane where the vehicle to be parked is located based on the number of the lane where the vehicle to be parked is located and a known position of each lane crossed by the shore crane relative to the one or more LiDARs;
clustering the real-time point cloud data based on the point cloud data of the lane where the vehicle to be parked is located, to obtain the set of point clouds for the vehicle to be parked.

11. The method of claim 1 , wherein said clustering the real-time point cloud dat

Assignees

• Beijing Tusen Zhitu Tech Co Ltd

Inventors

• Li Yiming
• Xiao Xu
• Jin Yuhe
• Cai Jinpeng
• Wu Nan

Classifications

• B60W2420/408

  Radar; Laser, e.g. lidar · CPC title

• G08G1/205 Primary

  Indicating the location of the monitored vehicles as destination, e.g. accidents, stolen, rental · CPC title

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title

• G01S17/42

  Simultaneous measurement of distance and other co-ordinates (indirect measurement G01S17/46) · CPC title

• G01S7/4802

  using analysis of echo signal for target characterisation; Target signature; Target cross-section · CPC title